The Undeserving Rich: “Moral Values” and the White Working Class

Abstract

White working-class citizens who vote for the Republican Party have been fodder for much political discussion and speculation recently, and a debate has arisen about the role that “moral values” played in the political decision making of this segment of voters. In this article, we defend a version of the moral values claim. We show that although the Republicans’ policies are unpopular, they are bundled with an overarching moral framework that is extremely resonant to this set of voters, and we use in-depth interviews to uncover this framework. A key feature of this framework, on which in the 2004 presidential election George W. Bush scored high and John Kerry scored low, is the appropriate attitude to wealth, which serves as an indicator for a candidate’s general moral philosophy and as a heuristic about whether the candidate will govern with working-class voters’ interests in mind. National Election Studies data support the argument that this was a key influence on the voting decision in 2004, even controlling for voters’ partisan identification.

Introduction

SUBJECT: I’ve noticed how hard it is to get back to work, I’ve put an ad in the paper … I’ve gone out and pounded the pavement …

S.H.: So what are you gonna do?

SUBJECT: What can I do? What can I do? My parents didn’t have the dough to send me to college … You can’t get anything now until you have an education. So what can I do to make things better so I can get my job before we lose our home?

S.H.: Do you guys have any ways of making extra money outside of your, outside of your formal job? You know, some people make things and sell things and that sort of stuff.

SUBJECT: No. Not really. We survive with a big garden … my husband hunts. And we eat squirrels. And, you know, rabbits and those kind of things, so I mean we can survive like that, if we, you know, be stuck on an island we’d make it.

Bush voter, October 2004

White working-class citizens who vote for the Republican Party have been fodder for much political discussion and speculation recently. Republican economic priorities seem to favor the wealthy at the expense of redistributive policies that would provide immediate benefits to larger segments of the population. Thus, support for the Republicans from less-wealthy segments of the population, like the voter quoted above, has puzzled observers. The scholarly literature attempting to explain this phenomenon has settled on the following six possibilities (Bartels, 2006; Brady et al., 2008; Burden, 2004; Campbell and Monson, 2005; Frank, 2004; Hillygus and Shields, 2005; Keeter, 2005; Kull et al., 2004; Langer and Cohen, 2005; Lewis, 2005; Lovett and Jordan, 2005; see especially Roemer, 2006:2–3).

• 1
  These voters like the economic policies because they believe these policies will help the general economy (and may ultimately help them, e.g., tax cuts for the wealthy might lead to more jobs for everyone).
• 2
  These voters like the economic policies because they believe that one day they, themselves, will become rich, and therefore will benefit from these policies.
• 3
  These voters like the economic policies because they agree with the general ideology behind them—that the rich become rich through hard work and should be rewarded for that, that private property should be sacrosanct—even if the policies do not benefit them personally.
• 4
  The economic policies are unpopular, but they are bundled with other issues that are popular with these voters, such as Republican positions on abortion, gay marriage, or foreign policy.
• 5
  The economic policies are unpopular, but the voters are unaware of them or misinformed about their true nature: they would not vote in this way if they were more informed.
• 6
  The economic policies are unpopular, but the voters are voting in this way because they prefer the “moral values” of the Republican Party.

In this article, we evaluate each of these arguments. We show that the first five arguments are not supported by the evidence, and we defend a version of the sixth argument. The “moral values” explanation received some early support from an exit poll taken during the 2004 presidential election, which found that nearly a quarter of voters chose “moral values” as their primary reason for choosing their favored candidate, and 80% of those who did so voted for Bush. In the immediate aftermath of the election, a wave of scholarly research dismissed the “moral values” argument. Fiorina (2004) pointed out that the wording of the exit poll question was too vague to allow for strong conclusions to be drawn (after all, fighting poverty can be seen as an issue of moral values) and Hillygus and Shields (2005) and others showed that issues that might be considered proxies for moral values were either statistically insignificant or had very small effects in predicting support for Bush. More recently, a second wave of research has suggested that there is something to the moral values argument after all. For example, Keeter (2005) presents other polling data showing respondents choosing “moral values” as important, Schuman (2006) defends the exit poll question, and Knuckey (2007) uses National Election Studies (NES) data to show that if “moral values” are understood more broadly, they do indeed show effects on voting.

Our findings are in accordance with this second wave of research. We show in this article that although the economic policies are unpopular, they are bundled with an overarching moral framework that is extremely resonant to this set of voters, and we use in-depth interviews to uncover this framework. A key feature of this framework, on which according to these voters George W. Bush scored high and John Kerry scored low, is the appropriate attitude to wealth—a key element of voters’ assessment of the candidates’ overall personality, which has been shown to be crucial in the particularly celebrity-based world of U.S. politics. This attitude serves as an indicator for a candidate’s general moral philosophy and, in particular, as a heuristic about whether the candidate will govern with working-class voters’ interests in mind. To elaborate this argument we draw on (and partially contend with) the work of Lamont (2000) on the moral boundaries of the white working class.

In the first section of this article, we show that Arguments 1 through 4 on the above list are not supported by NES data. In the next section, we report results from interviews that we conducted that contradict Argument 5. In the final section, we use these interviews to shed light on what respondents mean by “moral values.”

Explaining the vote: preferring republican policies?

NES data show that no matter how working class is defined, white working-class voters voted for George W. Bush in large numbers, and that they did not do so because they preferred Republican economic policies or the Republican stance on abortion. Arguments 1, 2, and 3 (see above) hypothesize that white working-class voters prefer Republican economic policies; the data below show that they do not. Argument 4 hypothesizes that white working-class voters prefer Republican social policies; the data below, as well as other studies, suggest that this is not the case.

Much confusion exists about the voting behavior of the white working class, most of it caused by disagreement about how precisely to define the working class. Reasonable dimensions of social class include annual income, educational attainment, and age.⁷Tables I–IV draw on NES data to disaggregate the effects of different dimensions of social class on voting behavior. One option is to define “white, working-class” voters as white voters with less than $50,000 in annual household income (roughly the bottom half of the income distribution). Bush earned 48.7% of the votes of this group compared to Kerry’s 49.7%; in the bottom third (roughly, those with less than $35,000), Bush earned 48.7% to Kerry’s 48.8% (see Table I). If social class includes an educational dimension, voters with college degrees should perhaps be excluded. If we look only at voters with less than $50,000 in household income and without college degrees, Bush earned 50.8% of the vote of this group, to Kerry’s 45.0%. If social class is a matter of self-identification, we can define our group of interest as white voters who call themselves lower class or working class. In that case, Bush’s advantage is 9.3 points. If we decide that class is best measured by education alone rather than by income, Bush had a 29.1-point advantage over Kerry among white voters without a college or associate degree. Finally, if we restrict “white working class” to those with a household income less than $50,000, without a degree, and between the ages of 25–65, Bush enjoyed a 21-point advantage. Although the exact level of support for Bush among white working-class voters depends on how the working class is defined, one thing is clear: a majority or near majority (over 48% in all cases) of the white working class voted for George W. Bush in 2004.

Table I.   Support for Bush and Kerry Among White Working-Class Voters, 2004 (% of Vote Given to Each Candidate), White Voters Only

Working Class Defined as	Support for Bush	Support for Kerry
Source: American National Election Studies data.
Less than $35,000 household income	48.7	48.8
Less than $50,000 household income	48.7	49.7
Less than $50,000 household income, no college degree	50.8	45.0
Define self as lower or working class	52.9	43.6
No college degree	62.9	33.8
Household income less than $35,000, no college degree, 25–65	Too few cases
Household income less than $50,000, no college degree, 25–65	57.2	36.2

Table II.   Support for Bush Among White Working-Class Voters Categorized by Region, 2004 (% of Vote Given to G. W. Bush), White Voters Only

Working Class Defined as Region of Country	Northeast	North Central	South	West	Entire Country
Source: American National Election Studies data.
Less than $35,000 household income	52.8	42.2	52.5	49.9	48.7
Less than $50,000 household income	48.5	43.3	54.2	50.3	48.7
No college or associate degree	41.9	59.8	71.5	72.7	62.9

Table III.   Support for Republican or Democratic Policies Among White Working Class (Defined by Income Less Than $35,000)

	Mean Voter’s Position	Mean Perception of Bush’s Position	Mean Perception of Republican Party Position	Mean Perception of Kerry’s Position	Mean Perception of Democratic Party Position
Note: Bartels (2006) methodology. Number represents voter’s self-placement on a scale of 1–7, and voter’s perception on scale of 1–7 of Bush’s position, Republican Party position, Kerry’s position, and Democratic Party position; standard deviation in parentheses; closer party/candidate in bold. Source: American National Election Studies data.
Spending and services	4.77 (1.67)	3.72 (1.67)	3.66 (1.52)	4.74 (1.52)	4.82 (1.45)
Defense spending	4.44 (1.60)	5.66 (1.40)	5.03 (1.34)	3.69 (1.67)	3.73 (1.48)
Women’s role	2.07 (1.63)	3.64 (1.78)	3.81 (1.58)	3.05 (1.53)	2.99 (1.45)
Jobs/standard of living	4.11 (1.88)	5.01 (1.53)	4.90 (1.51)	3.42 (1.52)	3.54 (1.50)
Gov’t asst. to blacks	4.79 (1.80)	4.53 (1.47)	4.54 (1.39)	3.56 (1.56)	3.62 (1.57)
Environment/jobs tradeoff	3.55 (1.61)	4.26 (1.72)		4.25 (1.49)
Gun control	2.05 (1.00)	2.76 (1.01)		2.63 (.98)
Interventionism/ diplomacy	4.03 (1.80)	5.71 (1.46)	5.37 (1.35)	3.18 (1.50)	3.39 (1.42)
Abortion	2.58 (1.12)	1.87 (.94)	1.95 (.87)	3.09 (1.02)	2.90 (.99)

Table IV.   Support for Republican or Democratic Policies Among White Working Class (Defined by Income Less Than $50,000)

	Mean Voter’s Position	Mean Perception of Bush’s Position	Mean Perception of Republican Party Position	Mean Perception of Kerry’s Position	Mean Perception of Democratic Party Position
Note: Number represents voter’s self-placement on a scale of 1–7, and voter’s perception of Bush’s placement on a scale of 1–7, Republican Party position, Kerry’s position, and Democratic Party position; standard deviation in parentheses; closer party/candidate in bold. Source: American National Election Studies data.
Spending and services	4.71 (1.59)	3.68 (1.65)	3.60 (1.55)	4.80 (1.53)	4.83 (1.49)
Defense spending	4.53 (1.53)	5.71 (1.31)	5.32 (1.30)	3.66 (1.60)	3.70 (1.46)
Women’s role	2.03 (1.57)	3.63 (1.79)	3.80 (1.59)	2.89 (1.48)	2.94 (1.44)
Jobs/standard of living	4.14 (1.84)	5.06 (1.51)	4.90 (1.47)	3.38 (1.48)	3.44 (1.46)
Gov’t asst. to blacks	4.74 (1.77)	4.67 (1.40)	4.68 (1.34)	3.52 (1.49)	3.49 (1.54)
Environment/jobs tradeoff	3.55 (1.61)	4.29 (1.67)		4.23 (1.49)
Gun control	2.03 (.99)	2.66 (1.00)		2.65 (1.03)
Interventionism/ diplomacy	4.02 (1.78)	5.79 (1.40)	5.38 (1.28)	3.07 (1.53)	3.29 (1.39)
Abortion	2.69 (1.10)	1.82 (.90)	1.92 (.84)	3.19 (1.00)	3.05 (.97)

Scholars have noted the strong regional bias of southern states for conservative Republicans, due either to racial resentment (Key, 1949; Valentino and Sears, 2005) or to economic development (Shafer and Johnston, 2006). Table II shows that working-class Republicanism was not restricted to the South in 2004. The table shows three ways of defining the working class, and the percent of support for Bush within each definition, categorized by Census region (again using NES data). Unfortunately, there are too few cases to reproduce this regional categorization using finer-grained definitions of the working class (e.g., less than $35,000 + no degree + ages 25–65), but the three definitions used here all show similar pictures. If the working class is defined by income, then with the exception of the North Central region, over 48% of working-class voters in all regions voted for Bush. If the working class is defined by education, then with the exception of the Northeast, a strong majority (59–72%) of working-class voters in all regions voted for Bush. The South’s Republican bias is high but not strikingly so: if working class is defined by education, then the highest support is in the West; if it is defined by income less than $35,000, the highest support is in the Northeast; only if the working class is defined by income less than $50,000 does the South rank highest in support for Bush, and then by less than 5 percentage points. This is not a robust demonstration that southern support is driving working-class support for Bush. Indeed, the picture is the opposite: in all definitions of the working class, in all income regions, support for Bush is over 40%, and in 9 of the 12 cases it is over 48%.

Tables III and IV show where voters place themselves, and where they place the Democratic and Republican candidates and parties, on several issues. Larry Bartels (2006) has shown that if “working class” is defined by education, working-class voters support Republican economic policies. But as we show here, if working class is defined by income, the pattern reverses: Tables III and IV replicate Bartels’s analysis using income rather than education to define working class, and show that working-class voters see themselves as closer to Kerry and the Democrats not only on economic issues, but in fact on all issues surveyed by the NES, with the exception of assistance to blacks. Bartels himself has explained why we should define the working class in terms of income rather than education: “Even in 2004, after decades of increasingly widespread college education, the economic circumstances of whites without college degrees were not much different from those of America as a whole. Among those who voted, 40% had family incomes in excess of $60,000; and when offered the choice, more than half actually called themselves ‘middle class’ rather than ‘working class’” (2006:205).⁸

If we follow Bartels’s method, then these data show that voters at the bottom of the income spectrum do not prefer Republican economic policies. It is also clear that abortion is not the issue driving voting behavior. The only issue on which white working-class voters are closer to what they see as the Republican position is the issue of governmental assistance for blacks, but this was not an issue in the 2004 election and, as we will show below, it does not remain significant when controlling for other factors.

Explaining the vote: information and misinformation

Many scholars argue that misinformation is the cause of this voting behavior: working-class Republicans are simply unaware of Republican economic policies. For example, Slemrod (2006) shows that the reason respondents support regressive reforms, such as a flat tax or a sales tax, is that they mistakenly believe that this will make high-income people pay more taxes. Kull et al. (2004) famously showed that misperceptions about the Iraq War were highly correlated with the decision to vote for Bush. Hacker and Pierson (2005a,b) argue that Republican politicians deliberately crafted the tax cuts of the early Bush Administration in order to increase public misperceptions about them. These scholars imply that properly informed voters would not have supported these policies.

This argument is based on a psychological model of information processing that scholars call “Bayesian updating”(Gerber and Green, 1999): in this model, decision making is a reasoned process of constantly updating a currently held belief, opinion, or hypothesis based on newly arriving evidence. Providing such voters with information about the Republicans’ policies would change their preferences. Bartels (2006:24), for example, compares the preferences of well-informed voters to demographically similar poorly-informed voters, and finds that more information correlates with more negative views about the tax cut. He writes: “If we take this cross-sectional difference in views as indicative of the effect of information on political preferences, it appears that the strong plurality support for Bush’s tax cut … is entirely attributable to simple ignorance.” But can cross-sectional difference in views be taken as indicative of the effect of information on political preferences? The desire to acquire the information, and the process of doing so, may make the demographically similar respondents in fact noncomparable. That is, it may not be the information that causes the difference between well-informed and poorly-informed voters, so much as exogenous preferences and the deliberative, information-seeking process itself. If this is the case, then information alone is not responsible for the change in preference, and it is not “simple ignorance” that is to blame so much as the social context that produces information and ignorance (Kunda, 1990; Lodge and Tabor, 2000; Lord et al., 1979).

To answer the question of whether and how information affects vote choice, we first conducted a survey of working-class Republicans. We chose counties in the researchers’ home states that were heavily Republican and majority white and had below-average incomes. We then selected the precincts within these counties that had voted most heavily for Bush in 2000. In North Carolina, we selected voters through publicly available voter registration records, selecting all registered voters who had voted for Bush in 2000. We took a 500-person sample from these voter rolls; the deletion of one duplicate record yielded a sample of 499 voters in North Carolina. Our approach varied slightly in Illinois because voting records are not made public in this state. We instead used precinct walk sheets to identify Republican voters. This generated 563 voters, all of whom we included in our sample. Our total sample consisted of 1,062 voters from both counties.

The survey asked questions pertaining to the respondent’s voting history and intentions, as well as basic demographic information, and requested permission to contact him or her for an in-person interview. Where possible, the language of the survey followed the language of analogous questions asked in the NES. The survey also asked two multiple-choice questions about political knowledge, one on foreign policy and one on domestic policy; the domestic policy question (relevant to the discussion here) was: “Some people say that rich people benefit the most from George W. Bush’s tax cuts, and other people say that the tax cuts benefit the average family more. Which do you think is true?” The possible answers were: “Rich people benefit the most from George W. Bush’s tax cuts”; “The tax cuts benefit average families more than rich people”; or “All families benefit about equally from the tax cuts.” The foreign policy question asked whether Saddam Hussein was responsible for the 9/11 attacks.

Of the 1,062 surveys we mailed, 12 were returned to us because of incorrect addresses, yielding a total sample of 1,050 respondents. Of these, 133 returned the surveys in Illinois and 134 in North Carolina; 267 in total, for a response rate of 25.4%. (Because this response rate is on the low end of the range of response rates generally considered acceptable for analysis [Dillman, 2000; Groves, 2006], we supplemented our analysis with analysis of NES data below. The findings from our survey are confirmed by the NES data, suggesting that any nonresponse bias in our survey does not affect the main substantive argument.) Of these respondents, 93 Illinois voters (70%) and 115 North Carolina voters (86%) reported having voted for Bush in 2000. Two people in Illinois (2.1%) and five people in North Carolina (4.3%) stated that they intended to vote for Kerry in 2004.

We found that political knowledge was a key correlate of voting for Bush (see Table V), even controlling for demographic characteristics: respondents who knew that Saddam Hussein was not responsible for 9/11 were extremely unlikely to vote for Bush (adjusted odds ratio .053), as were those who knew that the tax cuts were skewed toward the wealthy (adjusted odds ratio .123). This finding accords with research suggesting that information is a key predictor of vote choice.

Table V.   Estimate of Effect of Political Knowledge on Vote Choice, 2004 Presidential Election

	Model 1 Exp(B)	Model 2 Exp (B)
***p < .001. Odds ratio of voting for Bush in 2004. Hussein4 = “There is no evidence of a link between Saddam Hussein and the September 11 attacks.” TaxCuts1 = “Rich people benefit the most from George W. Bush’s tax cuts.”
Hussein4	.186***	.053***
TaxCuts1	.073***	.123***
Age	—	1.001
Education	—	1.047
Gender	—	.713
Income	—	1.104
Religious Attendance	—	1.135

However, to assess whether this correlation is evidence of causation (and in what direction), we then conducted in-depth “challenge” interviews designed to assess what would happen when these respondents were given correct information about the tax cuts. We interviewed all survey respondents who agreed to be interviewed, 51 people in Illinois and 33 in North Carolina. Of these, 41 in Illinois and 21 in North Carolina had voted for Bush in 2000 and 2004, and it is these 62 who are the subjects of the analysis below. Each interview took between 30 minutes and 2 hours, with most lasting approximately 1 hour. Interviewees were paid $40 for their participation. Interviews were transcribed and coded by the authors, as well as by undergraduate research assistants. We employed a semi-structured, expert-interviewer approach. Interviews were organized around an interview guide that specified the substantive topics to be addressed. These included a supplementary survey on the respondent’s political engagement and patterns of media consumption and questions on the respondent’s current political opinions and previous political biography.

The centerpiece of the interview was the employment of a challenge methodology, in which interviewers presented, and elicited participants’ responses to, widely available substantive criticisms of the Bush Administration. Since interviewees were Republican partisans, these criticisms presented them with explicit challenges to their political choices. These challenges addressed two topics—one domestic and one international—to which interviewees were asked to respond. First, interviewers offered evidence that President Bush’s tax cuts directly benefited the wealthiest Americans more than working classes. Second, interviewers provided evidence that Saddam Hussein was not involved in the September 11 attacks in the United States. These two topics were chosen because they were high-stakes issues with a correct answer, and scholars had shown that both issues were widely misunderstood during the 2004 campaign.⁹

For the tax-cut challenge, we showed interviewees two charts representing its distribution. The first, a bar graph, showed the average amount of money taxpayers in each economic quintile received back from the tax cuts, ranging from $250 for the poorest quintile to $7,740 to the wealthiest quintile. This graph further included a column for the top 1% of taxpayers, who received an average of $78,460 (see Fig. 1). The second, a pie chart, illustrated the percentage of the total tax cut each quartile received, ranging from 2% for the poorest quintile to 67% for the richest quintile (see Fig. 2). When presenting these two graphs, interviewers asked participants to give their reactions. The script’s specific wording, which interviewers occasionally modified slightly, was: “As you see in this chart, rich people are getting much larger tax cuts than other people. What do you think about that?” Interviewers prompted participants to describe their reaction as thoroughly as possible, encouraging the participants to engage with the information.

Figure 1.  Chart shown to respondents. Source: Wall Street Journal, August 13, 2004.

Figure 2.  Chart shown to respondents. Sources: Congressional Budget Office, Wall Street Journal, August 13, 2004.

In both cases, the point of the challenge methodology was to force partisans to engage with the most prevalent arguments against their positions. We did this not for political reasons, but in order to elicit rationales from voters for choosing their positions (had we conducted this study under the last Clinton Administration, e.g., we would have presented Democratic voters with evidence of Clinton’s infidelity or other challenging information, and tracked their rationales for minimizing, dismissing, or accepting it). Analytically, we sought to distinguish among the reasons voters used for selecting candidates. Such reasons might include agreement with the candidates’ positions; misunderstanding of those positions; agreement on other, more salient, positions; and preference for non-issue-related reasons.

Originally, only 35.5% of the sample knew the correct distribution of the tax cuts (see Figs. 3 and 4). When we presented those who did not know the correct distribution with information on the distribution of the tax cuts, one-third of these respondents decided that they approved of the tax cuts (22.6% of the total sample), and one-half decided that they disapproved of the tax cuts (32.3% of the total sample). (The remainder of those who did not already know the correct distribution—9.7% of the total sample—simply refused to believe the information.) However, none of these respondents changed their vote choice because of the information presented to them; those who disapproved of the tax cuts prioritized other issues or other reasons to vote for Bush.

Figure 3.  Knowledge and approval of Bush tax cuts among working-class Republicans.

Figure 4.  Knowledge of correct distribution of tax cuts among working-class Republicans (numbers in Fig. 3 do not sum to totals in Fig. 4 due to rounding).

This brings the total rate of approval of the tax-cut policy to 51.6% of the total sample, or over half (see Fig. 5). For the rest of the sample, while the information did lead them to disapprove of the policy, it did not lead them to change their vote preferences. A total of 38.7% disapproved of the tax cut but continued to favor Bush, even after those who had been misinformed were informed about the distribution of the tax cut. The remainder simply dismissed the contradictory information, refusing to believe that the tax cuts favored the wealthy. There were no major demographic differences between the group that knew the correct distribution and the group that did not. The group that approved of the distribution (including those who originally knew, and those who were informed in the interview) was composed of significantly more men than women: women were more likely to disapprove of the tax cuts, both those who knew the distribution, and those who were informed of it during the interview (cf. Brady et al., 2008).

Figure 5.  Approval of tax cut policy among working-class Republicans (numbers in Fig. 3 do not sum to totals in Fig. 5 due to rounding). (Total sample: Knowledgeable Voters + Non-Knowledge Voters After Being Informed.)

Thus, while almost all of our interviewees refused to change their preferences, they used two different strategies for doing so (in addition to the small minority that simply refused to believe the information): they either developed reasons for approving of the policy, or they prioritized other reasons for voting for Bush. Those who supported the tax cuts argued that the wealthy pay more taxes, so they should get more back, or that the rich deserved their wealth (the Democratic alternative to cut payroll taxes was not mentioned by anyone); that tax cuts for the wealthy will ultimately help them because the wealthy will create jobs; or that government wastes money anyway. Those who disapproved of the tax cuts prioritized other issues as more important to them than economic issues, or gave reasons other than issues for voting for Bush (e.g., his character, or their previous pattern of voting for Republicans). Very few respondents prioritized social issues (such as abortion or religion) in responding to the challenge; the modal response was to express disapproval of the policy, but to say “I’m just sold on Bush.”

I trust my family and they’re all gonna vote for him and, so, even though that he did have a tax, he did do tax cuts and everything, which would be against him, but I’d still vote for him, yeah.

Our analysis suggests that misinformation cannot be conflated with mistaken preference: correct information led the interviewees either to evaluate the information in ways congruent with their preference, or to minimize its salience in their decision making. Our contribution to the debate on information is thus to note that the majority of respondents were unaware of the distributional implications of the tax cuts; this lack of knowledge was significantly correlated with voting for Bush in 2004 and yet, when given evidence about the actual distribution, they resisted and dismissed it. If information alone were the cause of the preference, we would expect a pattern of more careful attention to the information and more careful weighing of it. Instead, our interviews suggest that the correlation between misinformation and voting Republican is better explained by prior preferences leading to resistance to contradictory information. But what is causing those prior preferences?

Explaining the vote: morality and the white working class

The final explanation that has been given for working-class Republican voting is that such voters vote Republican because of “moral values.” The question of whether “moral values” contributed to voting behavior in the 2004 election has, as noted, become a contentious one. First, as has been discussed above, it seems that abortion and gay marriage—the two issues that are widely held to have something to do with “morality”—did not have a strong or large influence on the electorate at large. However, as Lovett and Jordan note: “We believe that the emerging scientific consensus concerning the (non)impact of same-sex marriage and abortion on the 2004 election, rather than settling the ‘moral values’ issue, actually adds interest to the exit poll and survey results. Why did nearly half of Bush voters cite moral values as their top issue, if abortion and same-sex marriage—the most plausible policy issues for which ‘moral values’ may have been a stand-in—failed to have any independent effect on candidate choice?” (2005:166). Even if prolife and antigay marriage attitudes did not affect voting, there is still something to be explained: the fact that 22% of voters chose the phrase “moral values” as best explaining their voting behavior, and the fact that 80% of these chose George W. Bush. Keeter (2005:4) notes that the NEP poll was not the only one to find this pattern.

The Los Angeles Times exit poll found that “moral/ethical values” led the list in both 2000 and 2004. In 2000, it was checked by 35% overall, and 55% among Bush voters. In 2004, 40% picked it, including 54% among Bush voters. In both years, the list also included abortion (in 2000 as a separate item checked by 14%, and in 2004 in an item labeled “social issues such as abortion and gay marriage,” which was checked by 15%). The fact that “moral/ethical values” led the lengthy Los Angeles Times list in both years—one election with and one without an overriding national security issue in play—strongly suggests that the term captures an ongoing and strong concern among a segment of voters.

Scholars may have jumped the gun in concluding that the phrase “moral values” is equivalent to the abortion and same-sex marriage issues. But if not abortion and same-sex marriage, what, exactly, does “moral values” mean? Even critics of the exit poll note that “this phrase particularly resonates” (Langer and Cohen, 2005:745) with Bush voters. Why does it resonate, and what does it resonate with?

In this situation—a clear demonstration that a certain term (“moral values”) was meaningful enough for poll respondents to choose it as their primary motivation for voting a particular way, combined with confusion about the actual meaning of this term—inductive research is necessary. Thus, in this section we use our in-depth interviews to showcase our respondents’ understanding of morality and, specifically, one strong and recurrent moral evaluation that clearly works in favor of George W. Bush. We identified these frameworks in two ways: first, we identified spontaneous uses of the words “moral” or “morals” by the interviewees, and categorized what they associated with these words; and second, we coded interviews according to whether they made a moral evaluation—that is, one that gave a clear indication of a certain behavior or position being right or wrong¹⁰—even if the words “moral” or “morals” were not used.

In this section, we examine three ways in which a candidate’s moral values might be defined: as stance on policy issues; as commitment to traditional family life; and as personal traits. The first two have been mentioned in the literature, and while we find traces of them in our interviews as well, the evidence does not support seeing these as the main interpretation of “moral values.” Rather, our interviews and the NES data both suggest that candidates’ personal traits and comportment were what respondents were attempting to reference by the phrase “moral values.” Moral values, in other words, stood in for a mode of evaluating the presidential candidates as celebrities (Kurzman et al., 2007): individuals demonstrating symbolic qualities voters then use as cues to political preferences (see Norton, 1993; Rubenstein, 2008). This mode of reasoning has important racial and gender overtones (Bederman, 1995; Messner, 2007) alongside the more obvious class issues.

Morality as Stance on Policy Issues

Some of our interviewees did name specific issues as contributing to their belief that that the Republicans were more moral than Democrats, specifically, abortion, gay marriage, and stem cell research. Jim Tassel of Illinois focuses on the morality of abortion.

I would say abortion is, is strong. I believe, you know, you can’t kill for any reason and, and I think Bush has the best stance on that although I don’t agree totally with his stance because I think he supports it in some areas and I don’t at all. So, I think that, that would have to, probably be a number one issue of that sort, of a moral issue.

Chad Graves, a 31-year-old from North Carolina, supports Republicans for their stance on gay marriage.

I think it depends on morals and my belief is a biblical, and the biblical belief is that gayness is an abomination and God doesn’t like, I think it’s a moral degradation and I am glad the president took the stance he did because that’s what I would have wanted a representative to do. I have no concession for legalizing gay marriage for any reason.

Patrick Hill, a 53-year-old Illinois resident, ties stem cell research to abortion.

And, when it comes to moral issues, aside from his faith, things I feel strongly about, this stem cell research which I believe is, you know I know there’s a lot of people crying foul that they want to have it but what has to happen to have stem cell research a lot of it is aborted babies et cetera. And I don’t agree with that.

These issues have accurately been seen as referencing some aspect of morality during the 2004 election. However, as discussed above and shown below in the NES data, by themselves these issues do not explain the outcome of the election, and in our sample we also found cases of people who identified these as “moral issues” but disagreed with the Republican position on them.

Morality as Traditional Behavior

In addition to mentions of specific issues were mentions that identified morality as traditional behavior: respecting elders, making an effort to behave decorously in public, and especially marital fidelity. Ashley Rogers, a 54-year-old resident of Illinois, reflects on corporal punishment.

And I just see that today’s generations, this “time-out,” I’m sorry. [Laughs] There’s nothing wrong with takin’ a swat across this child’s rear end because their brain is not in their rear end.

Ashley’s example of disciplinary practice is interesting because it matches exactly George Lakoff’s popular explanation for the different “frames” that Republicans and Democrats draw on to evaluate politics.

However, issues like disciplining children are part of a larger whole. Consider the quote below from Sheila Johnson, a retired 64-year-old who lives with her husband in the trailer park on the edge of one of our working-class communities. We quote her at length to show the ways in which issues like abortion and gay marriage are easily incorporated into a much larger moral framework that includes approval of traditional family forms, respect for authority, and even sartorial choices.

I think we’re at a crossroads. Uh, for the moral fiber of our country. I’m not saying that Kerry isn’t a moral man, I do not mean that. But I think that President Bush has taken a stand for morality in our country, and if we don’t pay attention, it’s like uh, I’ve gone to church all my life, and I can remember growing up in church, we always dressed up to go to church on Sunday morning, you know? We wore our hats and our gloves. And then I go to church and there’s girls there in their blue jeans and their cut off tops and stuff and I think, all right, we’ve gone this far, what’s gonna, in ten years from now are they gonna come in their bikinis? … to me marriage is between a man and a woman. And they say well, if you say ok it can be two men, why not two men and a woman? When is that gonna come? … I had three sisters and two brothers. Three of us are divorced and three of us have been married over thirty years. Half and half. Fifty percent. … [sighs] What’s wrong? You know uh, and and, people, all wanting to have children so bad and yet, why don’t you give your baby up, let somebody else enjoy it, raise it? Why kill it? … I get on my grandchildren for calling our neighbors Jack and Anne. Now see that’s Mr. and Mrs. to you. That’s how I grew up, you called your older neighbors Mr. and Mrs. That’s a sign of respect. We don’t respect anything. We don’t respect, like they’ll say, well you know George did this or George did, that’s President Bush. That’s how, you should respect him. Even if you don’t agree with him. Respect him. I didn’t call, I don’t even know what Clinton’s first name is [laughs].

One surprise from our interviews is the frequency of reference to the immoral behavior of Bill Clinton, even 4 years after Clinton left office. Sixty-eight-year-old Anne Hawthorne says:

At first I liked [Clinton] and the more he was president I sometimes just wept because he was teaching our children such bad morals. And that’s just my opinion. But I think he did, was a good president. I really did, I voted for him. But then at last I was really gettin’ disturbed with him. I really was. I mean, emotional. For our children. I don’t know, that’s my opinion.

Even more striking is the comparison Sarah Belmont, a 44-year-old clerical worker, makes between John Kerry and Bill Clinton.

John Kerry reminds me so much of Bill Clinton …. just his mannerisms and everything. … Clinton had so many issues when he was in office you know, but with what was going on in his personal life that he met, that took the stage and took the focus away from what he was really there for. I, you know, I would try to stay clear from that. And that’s the impression I get from Kerry … he’s just, he just has the same air about him that reminds me of Bill Clinton. And I was not impressed with him when he went into office let alone what he went out with.

Evaluations of Clinton as not participating in traditional moral behavior seem to have colored other Democrats and, in particular, are blamed for teaching children poor personal morality (Eliasoph, 1998; Perrin 2006:98–99; Sarfatti-Larson and Wagner-Pacifici, 2001). George Bush, on the other hand, benefits from a perception as a family man. Forty-eight-year-old Thomas Robinson says:

Well I really, I think probably I do identify with him more because he’s a very much a family man, you can tell that he comes across, when he’s speaking a lot of times, you can tell he’s speaking from his heart, and that he’s a very much family man. He has a very great respect and love for his wife and his children and that shows in what he says. And so I think you know, a lot of that, that I think that’s with what all his judgments are made, through your character, and that’s where it comes from. So, I do like him a lot as a person I believe.

Jonathan Haidt and others have argued for the existence of different moral frameworks driving the left and the right (e.g., Haidt and Graham, 2007). Lovett and Jordan (2005) show that Bush voters are more likely to see the world in moral terms than Kerry voters (although their evidence for this is far from conclusive, as they only surveyed undergraduates). Knuckey argues that “moral values” should be understood as representing divergent positions on the “traditional-tolerant” spectrum, that is, as representing divergent answers to questions such as “This country would have fewer problems if there were more emphasis on traditional family ties” (2007:228); he finds support for this in the NES.

However, while we do find traces of this definition of moral values in our interviews, in the NES data this does not appear as a significant explanation of the outcome, as we will show below.

Morality as Personal Character Traits

A final aspect of what morality meant for our respondents was the candidates’ personal character traits as revealed through public accounts of their personal lives. Many observers have noted that Kerry came off as more aloof than Bush in the 2004 election. For our respondents, this difference was not trivial, and it was not simply an indicator of social distance: rather, it was an indicator of the candidates’ overarching character and philosophy. In particular, the way the two candidates—both wealthy men—handled themselves in relation to their wealth served as a key heuristic for our respondents, a clue to their moral fiber. Kerry was identified as being what we call part of “the undeserving rich,” whereas Bush’s wealth did not demean his character because he was a member of the “deserving rich.”¹¹ Sarah Belmont says:

Um, where [Kerry]’s coming from, you know, the person he’s married to, I think, money is a big thing in his eyes. That, I just don’t see a lot of solid values with him, um, George Bush I can see, I mean, he’s probably had his personal issues over the years too but I see him a lot more being able to confide in, um, more of a trustworthy person as far as his wife is concerned, I can see that solid relationship there which I think brings a lot to the table. Um, and not to say that that man doesn’t have money either, but I say it more in an American family way than Kerry. So um, that’s the impression I get when I see them, and the spouses and so forth.

For Sarah, there is an “American family way” to be rich, and there is a way to be rich that involves not having solid values, believing that “money is a big thing.” The markers of the undeserving rich are an excessive appreciation of and display of money, and for John Kerry, that is indicated by his choice of spouse.

For Sheila Johnson, Kerry’s wealth did not accord with his claims to favor populist policies.

I appreciated that uh, [Bush] cut taxes and it always amazes me, they always say that he cut taxes for the rich and yet when I hear Kerry talk about it I say, did you take those tax cuts? ‘Cause you’re rich. And it’s kind of ironic to hear him talk about us poor middle class when he has five homes and all these cars and it just doesn’t jibe to me. You know it doesn’t. How can he know how we feel when he’s not one of us either?

In this quotation, a criticism of Bush—that his tax cuts went to the wealthy—becomes a criticism of Kerry through the fact of Kerry’s personal wealth. Although both Kerry and Bush are acknowledged to be wealthy, Kerry’s wealth makes his claim to prefer policies for the middle classes suspicious.

Similarly, Michael Smeed, a 50-year-old who gives “disabled” for his occupation and says that he would vote for Hillary Clinton if she were running, also brings up Kerry’s wealth in response to the evidence of the distribution of the tax cuts.

INTERVIEWER: … I want you to take a look at these two charts here. It just shows kind of what the breakdown of who’s getting the tax cuts. And uh, as you can see, most of the tax cuts are going to some of the richer people. What do you think about that?

MICHAEL SMEED: Mmmm, I don’t like it but they got more money. Yeah, if I was rich, that would be great.

INTERVIEWER: Right, right.

MICHAEL SMEED: [long pause] Hell of a big difference.

INTERVIEWER: Yeah, yeah, the other one is like a bar that kind of shows um …

MICHAEL SMEED: Wonder what Kerry’s would have been.

When shown the distribution of the tax cuts William Smith, a self-employed builder, says:

Okay. So I’m guessing, you say top 1%. You know, is that fair? … I wish everybody could be right here in the middle and unfortunately it’s never going to be that way. You’re going to have the Heinz and Rockefellers and all those others that are gonna be in the top end and you’re gonna have the poor people and I have no idea how to ever change that.

It is not clear in these last two examples whether the interviewees are bringing Kerry up rhetorically, to deflect criticism of Bush, or whether Kerry (or “Heinz”) are simply the first names that spring to mind when the discussion turns to wealth.

Indeed, for some of our respondents, the very mention of Kerry’s name brings condemnation of his wealth. Consider Timothy Hawthorne (husband of Anne), the father of a family that, unusually, consists of both Democrats and Republicans:

INTERVIEWER: Hmm. So what do you think of John Kerry?

TIMOTHY HAWTHORNE: Who?

INTERVIEWER: John Kerry.

TIMOTHY HAWTHORNE: Well, to tell you the truth, I don’t know if he’s a bag of wind or [laughs]. I hear so many things he’s gonna do and I, you know, I just can’t, I just don’t, I don’t really believe him, I don’t. I don’t really believe him. I believe his wife is a Heinz and she’s got more [coughs], well, he’s got about seven more mansions than Bush has got, he got more money, but, uh, I don’t know why.

Unlike many of our respondents, Dave Wilson talks regularly with Democrats. Dave is knowledgeable about some key political issues, such as the absence of a link between Saddam Hussein and 9/11. But when asked at the end of a long-ranging interview whether he could give one reason why a Kerry voter should vote for Bush, he says only:

It just, I couldn’t put it down to one thing. If I’m gonna have to put it down to one thing I’d say probably too much money [laughs].

To this interviewee, the fact of too much money should be enough to convince a Kerry voter to vote for Bush.

Evaluations of Kerry’s wealth were tied to the persona that he projected on television.

I’m not real, real keen on, on people who present the persona that they’re just a little bit better than you are. A little bit smugger. That’s the way he came across to me on the TV. Uh, it’s uh, but I think that’s a New England part of it. You know, if you’ve got a lot of money, and you’ve been a senator for twenty plus years, you do get that kind of aura about you, I realize that, you know you just got to. You know. But I’m not fond of it.

George Bush’s persona, however, shows that he has not let his money taint his values.

Oh yeah, I think that he’s all right. He’s just a down-to-earth person, you know? I don’t think he let money go to his head. I think they’re still as plain as everybody. I think he’s just a regular cowboy, a cowboy rancher.

Kerry seems aloof, whereas Bush seems like he is someone it’s possible to talk to.

I think I could talk to probably the President a lot easier than I could with Kerry. He seems like he is um, I don’t know, Kerry seems like he might be what you call upper level and a little snooty, and I don’t know. But I never met him so I can’t really pin that him but that’s what I kind of perceive him as, where Bush is just Bush, so.

In all, 16 of the 62 Bush voters (25.8%) mentioned some variant of this theme. All these mentions were spontaneous.

But why were these distinctions in appearance so important to these voters? And what does being a cowboy rancher have to do with morality?

For help in clarifying these questions, we turn to two recent ethnographies of the white working class, Kefalas’s Working Class Heroes (2003) and Lamont’s The Dignity of Working Men¹² (2000). Both these studies—conducted and published before the 2004 election—converge on the finding that many working-class respondents argue that morals are a more important marker of worth than socioeconomic status. As Lamont explains it, this focus on morality is a “counter-ideology” to the American dream and its assumption that prosperity is a sign of virtue and a reward for a strong work ethic. By privileging personal morality over economic success, workers create a hierarchy in which it is possible for them to be on top: “Their counter-ideology, which revolves around the simple idea that people on top are no better than themselves, contests one of the central tenets of the American dream by questioning a posited link between social position and merit. In this respect, they appear to be less under the spell of the market logic of evaluation than many observers suggest” (2000:114). Lamont finds this morals-over-money evaluation among one-third of her respondents. But not all rich people are condemned. There are some rich people, Lamont writes, who are seen as recognizing and appreciating the dignity of workers. Lamont gives the example of a wealthy client who invited one of her respondents, an electrician, to a party.

The electrician respects his client because he recognizes the value of workers’ skills and accords them dignity. He does not deny their humanity on the basis of status differences. He signals all of this by inviting them to a party. For this electrician, the ability to recognize workers’ dignity is the standard by which the worth and quality of a person is measured, and the client’s ability to do this is explained by his working class roots … By subordinating social status to what they perceive to be the “real” value of a person, workers create the possibility of locating themselves at the top of the hierarchy. (2000:111, emphasis in original)

Meanwhile, rich people who are morally bankrupt can be identified by distinct identifiers: the way they talk and dress, and the meaning that they seem to place on having money. One of Lamont’s respondents, when asked how he identifies the rich people he does not like, says: “the way they dress, the way they talk, the way they treat you” (2000:109). Another says: “I was always taught one thing. When an educated person comes in, he should always speak to the level of the person in the house. [You] don’t go in a person’s house and look down on people.” The key criterion distinguishing between the deserving rich and the undeserving rich is that the deserving do not look down on others—that is, the deserving recognize the workers’ own criteria of moral evaluation. And the key method of identifying the undeserving is by their mannerisms, specifically, a certain way of talking and the value that they seem to place on money.

This is seen clearly in the answer one of our respondents gives when asked to imagine himself at a party with the two candidates. Adrian Healy comes from a military family and disparages Kerry’s military service; he says he isn’t voting for Bush, but rather against Kerry. When asked how Bush and Kerry might behave at a party, he responds:

ADRIAN HEALY: From what I’ve seen of the two on television, George Bush would probably talk to me. Kerry wouldn’t. He’s too high and mighty to talk to a normal human being.

INTERVIEWER: Gotcha. Who do you think he’d be talking to instead?

ADRIAN HEALY: Well probably one of his advisers or somebody with money. He wouldn’t be, I know he wouldn’t be talking to me.

However, although Lamont sees this morals-over-money evaluation (that there is no link between wealth and worth) as contesting the ideology of the American dream, there is a key way in which this evaluation ends up supporting the ideology of the American dream: if morals are more important than money, then a candidate’s proposals for economic redistribution are less important than the candidate’s ability to recognize the human worth of those at all levels of the economic spectrum.

With a grandson in jail for 20 years, the issue of the morals that children are learning is not an abstract one for Anne Hawthonre. She struggles with the issue while considering the presidency of Bill Clinton.

I think he taught our children that it’s ok to stand in front of the cameras and lie. And our children are going downhill right now. I see it. The morals are changed and they haven’t got that good solid ground they’re standing on. And, you know, I think he did keep our jobs good and our economy, all that, the bonds, stocks going up, but is money everything? It’s important to live, I just got done griping about my medicine, but I think your morals are just as important as, you know, your foundation and everything else.

Anne’s “gripe” about her medicine is that she has to choose between buying her medicine and paying her phone bills; she is being helped by a prescription drug program in Illinois. She appreciates what she sees as Bill Clinton’s ability to protect jobs, and she did vote for Clinton, but she is deeply troubled by his behavior. Meanwhile, she has been confused about the Bush tax cuts, and does not have a clear response to them when shown the chart.

This is one of the things I was a little mixed up on. But, the morals I stand up for, you know, this has a little concern for me but not near as much as life. I’m the poor people, so. And I’m surviving. I eat good. You know, sometimes it’s hard, but.

Our findings support Lamont’s findings about a “morals over money” norm among the white working class, but where Lamont argues that this norm underpins an egalitarian critique of the American dream, our interviews suggest that this norm can also be used to discount the importance of material redistribution. If morals are more important than money, then the personal morals of political figures—including whether they seem to see themselves as better than “normal human beings” because of their wealth—are more important than socioeconomic policies of redistribution.

Although the NES data include too few working-class respondents to allow an evaluation of this hypothesis among this population specifically, the argument holds among white respondents in general: when put into a regression with the usual battery of demographic variables (gender, age, education, union household, income, subjective class position), financial and sociotropic beliefs (whether the respondent is personally better off and whether the national economy is better off), and political preferences and positions on issues (liberal-conservative ideology, party identification, approval of war, defense spending, government responsibility for jobs, government assistance to blacks, environment vs. jobs tradeoff, death penalty, guns, importance of religion), the “moral values” variables are among the best predictors.¹³ However, this is not moral values as policy issues: abortion and gay marriage are not significant. And it is not even moral values as general moral climate or preference for traditional family forms: these variables are also not significant. But the variables that represent morality understood as character traits are highly significant, and large in magnitude. We have chosen two specific variables to represent the themes that we saw in our interviews: whether respondents agree that the trait “cares about people like you” describes the candidates well, and whether respondents agree that the trait “moral” describes the candidates well.

Table VI presents the full regression results. The only six variables that are significant in predicting a vote for Bush are, in decreasing order of significance: (1) disagreeing that the trait “cares about people like you” describes Kerry; (2) approving of the way Bush is handling the Iraq War; (3) Republican Party ID; (4) agreeing that the trait “cares about people like you” describes Bush; (5) agreeing that the trait “moral” describes Bush; and (6) higher income. It is interesting to note that the question asking whether the trait “moral” describes Kerry is not significant: while the NES respondents do describe Kerry as not “caring” about voters like them, like Sheila Johnson above they are hesitant to actually call him immoral. But they are willing to call Bush moral, and they are particularly influenced by their opinion on whether the two candidates care about people like them.

Table VI.   Logistic Regression Predicting 2004 Presidential Vote Choice (White Voters); Dependent Variable is Vote Choice (Kerry = φ, Bush = 1)

	Model 1		Model 2
	b	SE	b	SE
*p < .10; **p < .05; ***p < .01. Source: National Election Studies data.
Constant	−19.918*	10.488	−6.698***	1.927
Party identification	1.524**	.597	.959***	.229
Liberal-conservative self-placement	.160	.301	—	—
Government aid to blacks	−.325	.417	—	—
Government guarantee jobs	2.94	.318	—	—
Defense spending	.405	.503	—	—
Iraq War	5.183***	2.020	2.975***	.716
Environment/jobs	.625	.417	—	—
Death penalty	.509	.666	—	—
Gun control	.221	1.179	—	—
National economy	.038	.482	—	—
Personally better off	−.584	.841	—	—
Gender	−1.076	1.417	—	—
Income	.392*	.201	.121*	.064
Education	−.042	.561	—	—
Age	.023	.042	—	—
Union household	.252	.332	—	—
Subjective class	−.494	.531	—	—
Religion important	.045	.399	—	—
Religious denomination	.064	.248	—	—
Abortion	−1.512	.940	—	—
Gay marriage	.354	.751	—	—
Moral climate	1.247	3.217	—	—
How much moral climate	−.515	1.853	—	—
Traditional family	.929	.791	—	—
Bush moral	2.882**	1.400	1.073**	.446
Bush cares	3.759**	1.482	2.723***	.724
Kerry moral	−.498	1.281	—	—
Kerry cares	−4.242***	1.415	−3.161***	.780
N	378		378
Cox and Snell pseudo R^2	.71		.69
Percentage correctly classified	97.9%		96%

Thus, one of the keys to white working-class behavior in the 2004 election is the perception of whether or not the two candidates seemed to care about voters. This perception was shaped by how the candidates handled themselves in relation to their wealth: Kerry’s perceived relationship to his wealth was read to mean that he considered himself superior to his fellow citizens. This is what voters meant when they cited “moral values” as the reason for their vote.

In his classic ethnography Learning to Labour, Paul Willis (1977) argued that working-class boys in Britain remain in the working class not because they are duped into accepting their class position, but precisely because they do not accept their class position. Their strategy of resistance is to reject school, the main institution representing the status quo in their world, which ironically condemns them to their class position. Similarly, many of our working-class respondents do not buy into the American dream ideology with its equation of wealth and merit: as Lamont points out, they are rejecting this ideology and its emphasis on material achievement in their efforts to maintain an alternative definition of morality and human worth. But their strategy of resistance is to emphasize nonmaterial factors and values as central to human worth, which leads them to downplay the need for policies of material redistribution and to look instead to the perceived personal qualities of candidates (qualities that can be manipulated by sophisticated political handlers). Their defense of a nonmaterial definition of human worth ironically reinscribes their class position. This is not the only factor affecting their vote choice, but it is a curious and counterintuitive theme that may partly answer the perennial question of why some members of the working class resist material redistribution, including in other countries in other periods of history.

Conclusion

We have argued here that “moral values” did play an important role in the white working-class support for George W. Bush. The policies of the Republican Party were unpopular among this segment of voters and cannot explain the votes. “Moral values” do explain the vote, but morality understood as character traits rather than as stance on policies or as traditional behavior. Other analysts (e.g., Bederman, 1995; Berman, 1970; Hochschild, 1995; Lamont, 2000) have demonstrated the extent to which this personalized view of politics is intimately intertwined with extant social inequalities.

Wealth, in and of itself, does not threaten a candidacy. But “undeserved” wealth does. Our respondents perceived Bush as part of the “deserving rich”—those whose mannerisms suggest that they are able to see the true human worth of everyone—while they saw Kerry as part of the “undeserving rich”—those who believe their elevated social position makes them superior to others, or who see money as an end in itself. This made the respondents reluctant to believe that Kerry could govern with their interests in mind. This heuristic led them to vote for Bush despite disagreement with his policies, and to resist information that would contradict the preference established through this symbolic judgment about morality.

Footnotes

• 7

   Including age makes sense because younger voters may be in the process of educating themselves on their way out of the working class, so including them may overstate the Republican vote for the working class; and older voters may have spent a lifetime in the middle or upper classes before ending up on pensions, so including them may also overstate the Republican vote for the working class. Bush’s advantage is greater still if we restrict the ages to 30–65, to exclude those who may be working toward graduate degrees.

• 8

   In an earlier version of his paper, which is a critique of Thomas Frank’s book What’s the Matter with Kansas, Bartels used income to define working class; Frank responded that the working class should be defined by education (Frank, 2005), and Bartels gamely obliged in the final published paper by using education as the index. Ironically, although Frank demanded that education be the determinant of working class (since at first sight that seemed to support his hypothesis), in fact it turns out that income is a better index of the puzzle that Frank wants to investigate.

• 9

   This article presents only the results of the tax-cut challenge. The foreign-policy challenge is analyzed in a separate paper (Prasad et al., 2009).

• 10

   For example: “Uh, Clinton would, he could smile at you, he was very uh, he made a very appealing appearance. But he was one way and then he was another way, um, when he said ‘I did not have sex with that woman’ then he lied right to all of us when he did that. How could you respect a man, a person like that?”

• 11

   These terms allude, of course, to scholarly categorizations of the ambivalent attitudes Americans (and others) hold toward the poor. We note here that a similar ambivalence can be identified in attitudes toward the wealthy.

• 12

   Anthropologists (e.g., Hartigan, 1999) and historians (e.g., Sugrue, 1996) have examined everyday life in such neighborhoods and the history of these neighborhoods, particularly in association with “whiteness studies,” but their theoretical concerns are distinct from those treated here.

• 13

   With some exceptions, we have followed Knuckey’s (2007) arguments about which variables to include.